基于机器视觉的红枣特征提取

Feature Extraction of Jujube Based on Machine Vision

红枣作为我国的特色优势农产品,受到各地的高度重视。针对我国红枣品质检测仍停留在靠人工进行识别判断的现状和机器视觉技术在水果品质检测中的广阔应用前景,研究了利用机器视觉技术精确检测红枣特征的方法。利用机器视觉系统首先获取红枣的图像,对获得的红枣图像进行图像分割,得到红枣轮廓图像,之后从红枣轮廓图像中提取红枣的边缘,最后从红枣边缘图像中提取红枣的面积、位置、周长等特征。实验结果表明,该方法检测速度快,正确率高,适用范围广,能够满足红枣自动级要求。

Jujube is the characteristic agriculture product of our country,and has attached great importance.Jujube quality detection mainly relied on artificial method.On the other hand,machine vision technology had board application prospects in fruit quality detection.In this paper,the method of jujube feature extraction was studied.Firstly,jujube image was captured,then image segmentation,finally extracted the feature of area,location,girth and so on.The method had the advantage of high speed,high correct rate and wide application range,and satisfied the requirement of jujube automatic classification.